"storage not initialized" when rotating local drives + feedback on 2 local backup strategies

Support
#1

Hello,

I’m finally taking the plunge into Duplicacy backing up my Unraid NAS shares. I plan on doing an “unconventional” offsite strategy by backing up my irreplaceable data to an external drive and storing the drive at my office under lock & key. I would be doing this once a month rotating the two drives so I would always have a month old copy of my data off-site.

I read a number of forum posts (https://forum.duplicacy.com/t/back-up-to-multiple-storages/1075,https://forum.duplicacy.com/t/same-storage-multiple-hdd/3286) but I’m running into an issue I can’t get past.

Scenario A
Single NAS Share → onsite_backup–(COPY)–>offsite-backup(usb)

  1. I setup 2 storage destinations (both are unassigned drives not part of the NAS array).
  2. I backed up a single NAS share to the onsite_backup desitination then w/ scheduler I use COPY to sync it to my offsite-backup destination (the drive I’ll rotate once a month)
  3. To test, I unmounted the offsite-backup drive, and mounted a second drive with the same path.
  4. Re-running the schedule, I received the following error:
Running copy command from /cache/localhost/all
Options: [-log copy -from onsite-backup -to offsite-backup]
2023-01-30 21:53:42.391 INFO STORAGE_SET Source storage set to /backup-mounts/downloads/Backups
2023-01-30 21:53:42.395 INFO STORAGE_SET Destination storage set to /backup-mounts/ext-backup
2023-01-30 21:53:42.395 ERROR STORAGE_NOT_CONFIGURED The storage has not been initialized
The storage has not been initialized

Scenario B
Single NAS share → ext-backup destination

  1. I also tried to create a single backup destination (ext-backup). I completed a single backup, I then unmounted the drive and mounted another drive w/ the same path, but it failed with a similar error:
Running backup command from /cache/localhost/0 to back up /shares/Client Photos
Options: [-log backup -storage ext-backup -threads 1 -stats]
2023-01-30 22:38:01.308 INFO REPOSITORY_SET Repository set to /shares/Client Photos
2023-01-30 22:38:01.309 INFO STORAGE_SET Storage set to /backup-mounts/ext-backup
2023-01-30 22:38:01.309 ERROR STORAGE_NOT_CONFIGURED The storage has not been initialized
The storage has not been initialized

So it’s easy enough to see that it doesn’t like the second drive since it’s not initialized (even though it’s the same path). I’m not sure how to initialize the drive without creating another destination.

Regarding local backup strategies: I’m thinking Scenario A is overkill, but using the copy method seemed best practice. I’d prefer having one less destination and do Scenario B if possible. I know I can get this working, just need advice from the wizards.

Best Regards,
Richard

#2

Don’t do dance with drives, you will lose data.

Imagine your first drive and the nas succumb to the power surge and your other drive in the safe fails to spin up, as they sometimes do, for no apparent reason, or develops bad sectors/rot. (2-3 years is all it takes), that you won’t notice until the restore time.

Not to mention you have expensive manual steps in your workflow (exposing you to Murthy law driven mishaps) and stressing drives moving them around all the time does not improve reliability.

Irreplaceable stuff shall go automatically over the network to remote storage. With no manual steps. That can be your own redundant nas at work, with redundant checksumming filesystem, or commercial storage provider, such as Amazon AWS S3.

This absolves you from manual steps and provides actual durability for your data.

Do not entrust your data to a single hard drive, they are replaceable, unreliable commodities, that don’t guarantee that you will read from the drive tomorrow what you wrote today.

1 Like
#3

In both scenarios, you simply need to initialise the second storage.

Scenario B still involves two storages, so you’ll need to initialise them separately with Duplicacy. In the same way as you did the first. (There’s a green plus (+) button at the bottom left of the Storage page.)

As you’re exploring copy anyway, I’d recommend you make them copy-compatible, thus the best way to initialise is to temporarily mount both drives and add the second, ensuring to tick the ‘Copy-compatible with’ option:

image
image537×564 17 KB

You’ll have two physical (and two logical) storages so keep in mind schedules, drive letters / mount points, and the fact Duplicacy doesn’t detect which drive is connected.

External hard drives are perfectly fine to use a backup destination - ignore anyone that says otherwise. So long as you have other copies (i.e. 3-2-1). However, losing a month’s worth of data isn’t ideal, and I’d highly recommend you use it in conjunction with live(!) off-site backups - to automate the process - and this is where copy comes into play. You could still, for instance, rotate drives for a 4th copy.

1 Like
#4

Thank you for the feedback @saspus and feedback/help @Droolio . I definitely gave cloud storage some deep thought but it’s not in the cards for me atm. My 3-2-1 backup strategy is enough for my needs.
Photos on Unraid array (source)
Photos on internal non-array storage (second copy)
Photos on offsite external drive (offsite copy rotated every month)

@Droolio when you mentioned

You’ll have two physical (and two logical) storages so keep in mind schedules, drive letters / mount points, and the fact Duplicacy doesn’t detect which drive is connected.

I’m not certain how to schedule the jobs since it seems like there will always be one job failing (the storage destination that’s not attached) since it looks like I can’t set a different schedule every other month. This is how it’s currently setup. Do I just have to let one job fail if I’ll be rotating disks?

image
image2026×656 65.4 KB

#5

Unfortunately yes with this setup…

Depending on your procedure for swapping out drives, perhaps you could put the copy jobs into separate schedules and pause (done by unticking all the days in the week) and unpause as needed? Bit of a faff I know.

There’s an alternative setup, but I was reluctant to suggest it as I’m not sure it’d work reliably. I’d definately do further research to see if it’s feasible. (I’m sure there’s been discussion about this scenario, but I can’t find it as yet.)

This is where you have just one logical storage (in Duplicacy Web) and you do indeed just swap the physical drivers over, making sure the mount point is the same. As your physical storages are initialised and contain the same snapshot ID, backups should work - even incrementally.

Potential issues may arise due to the fact you’d be using the same snapshot ID and the local cache may not be in a consistent state. Thinking about it, I don’t think this would cause issues, but it’s worth double-checking. :slight_smile: Perhaps @gchen can tell us?

Edit: Okay I found this:

#6

In the same post I see towerbr is accomplishing rotating the disks. @towerbr can you go into your setup a little more if you don’t mind since I’m also looking to rotate disks offsite? Duplicacy gave me a “Disk not initialized error” when I swapped disks that had the same mount path.

#7

So long as you’ve previously initialised the physical storage with Duplicacy (you can delete the logical storage after), which results in a folder structure starting with a config file, and eventually a snapshots and chunks directories… you should be able to swap the HDD out with another with the same structure.

#8

This is an old post, from another life :wink:, today I use only cloud.

But what I did - using CLI - was simply considering the two storages/HDDs as independent, I never made copies between them.

#9

Copying the config from one drive to the other did the trick for me. I’m doing a test backup to an internal-drive (storage destination labeled “int-backups” and am able to run a copy operation to a single storage destination (“ext-backup”) which is working even when rotating disks w/ the “ext-backup” mount point. I did a successful test restore from both drives (w/ RSA encryption). Still have a number of things to do before I do full backup. If/when I go the cloud route, I plan on just doing a “copy” job from my main “int-backup” drive. Thank you everyone :slight_smile: